Re: What's cooking in git.git (Nov 2010, #02; Wed, 17)

[Jakub Narebski <jnareb@gmail.com>]

On Wed, 24 Nov 2010, Junio C Hamano wrote:
> "J.H." <warthog9@kernel.org> writes:
> 
> > While I don't disagree, being able to support other caching systems
> > would be nice, we have this now, it's tested and it works.  I'd argue
> > this is a step 1, step 2 case at this point.
> 
> I would have to agree; getting some cache layer that is known to work (you
> are battle testing your updates on a busy site of yours as you develop,
> right?) early, quick and dirty with minimum changes, is more beneficial
> than waiting for a large rewrite thats "gets it perfect this time".

I also agree, that's why I'd like to go the "minimal fixups" route, where
those minimal fixups are about caching being robust (loading cache.pl,
passing minimal tests), not penalizing non-cached case (moving check for
$caching_enabled out of cache_fetch), and make it possible to move to
better solution / other caching engines in the future while maintaining
backward copmatibility (configuration is API; $caching_enabled, perhaps
also $cache (is 'cache.pl') and %cache_options).

> > I'm currently working from on top of Jakub's last tree, though I've got
> > some questions about his reasoning on a few things now that I've been
> > digging into it.
> 
> Thanks both for keeping the ball rolling.
> 

-- 
Jakub Narebski
Poland